Alterations in Glucose Metabolism in the Elderly ... - Diabetes Care

2 downloads 0 Views 937KB Size Report
OBJECTIVE — To determine the alterations in glucose metabolism in elderly pa- tients with NIDDM. RESEARCH DESIGN AND METHODS— We studied 9 ...
O R I G I N A L

A R T I C L E

Alterations in Glucose Metabolism in the Elderly Patient with Diabetes GRAYDON S. MENEILLY, MD KEITH DAWSON, MD

DANIEL TESSIER, MD

OBJECTIVE — To determine the alterations in glucose metabolism in elderly patients with NIDDM. RESEARCH DESIGN AND METHODS— We studied 9 healthy elderly control subjects (73 ± 1 yr of age; body mass index 25.7 ± 0.4 kg/m2) and 9 untreated elderly NIDDM patients (72 ± 2 yr of age; BMI 25.9 ± 0.5 kg/m2). Each subject underwent a 3-h oral glucose tolerance test (40 g/m2); a 2-h hyperglycemic glucose clamp study (glucose 5.4 mM above basal); and a 4-h euglycemic insulin clamp (40 mM • m2 • min" 1 ). Tritiated glucose methodology was used to measure glucose production and disposal rates during the euglycemic clamp. RESULTS— Patients with NIDDM had a higher fasting glucose (9.3 ± 0.3 vs. 5.1 ± 0.1 mM in control subjects vs. NIDDM patients, respectively, P < 0.001) and a greater area under the curve for glucose during the OGTT (16.0 ± 0.6 vs. 6.7 ± 0.3 mM in control subjects vs. NIDDM patients, respectively, P < 0.01) than the healthy control subjects. During the hyperglycemic clamp, patients with NIDDM had an absent first-phase insulin response (112 ± 6 vs. 250 ± 31 pM in control subjects vs. NIDDM patients, respectively, P < 0.01), and a blunted second-phase insulin response (159 ± 11 vs. 337 ± 46 pM in control subjects vs. NIDDM patients, respectively, P < 0.01). Before the euglycemic clamp, fasting insulin (99 ± 5 vs. I l l ± 10 pM in control subjects vs. NIDDM patients, respectively) and hepatic glucose production (11.8 ± 0.7 vs. 11.5 ± 0.5 (junol • kg 1 • min l in control subjects vs. NIDDM patients, respectively) were similar. Steady-state (180-240 min) glucose disposal rates during the euglycemic clamp were slightly, but not significantly, higher in the normal control subjects (36.5 ± 1.1 vs. 33.1 ± 1.9 |xmol • kg" 1 • min in control subjects vs. NIDDM patients, respectively, NS). CONCLUSIONS— We conclude that NIDDM in nonobese elderly subjects is characterized by a marked impairment in insulin release. This may be attributable to the toxic effects of chronic hyperglycemia on the (3-cell. When compared with age-matched control subjects, the NIDDM patients showed no increase in fasting insulin or hepatic glucose production, and insulin resistance was mild.

From the Department of Medicine, University of British Columbia, Vancouver, British Columbia; and the Department of Medicine, University of Sherbrooke, Sherbrooke, Quebec, Canada. Address correspondence and reprint requests to G.S. Meneilly, MD, Room G 433, Jean Matheson Pavilion, University Hospital, Shaughnessy Site, 4500 Oak Street, Vancouver, British Columbia, Canada V6H 3N1. Received for publication 8 March 1993 and accepted in revised form 3 June 1993. NIDDM, non-insulin-dependent diabetes mellitus; BMI, body mass index; OGTT, oral glucose tolerance test; HGP, hepatic glucose production; HGO, hepatic glucose output; CHO, carbohydrate; NDDG, National Diabetes Data Group; CV, coefficient of variation; R,, rate of glucose appearance; R^ rate of glucose disappearance; BP, blood pressure; sBP, systolic blood pressure; dBP, diastolic blood pressure; RIA, radioimmunoassay; AUC, area under the curve; WHR, waist-to-hip ratio; ANOVA, analysis of variance; FBG, fasting blood glucose; GDR, glucose disposal rate.

DIABETES CARE, VOLUME 16, NUMBER 9, SEPTEMBER 1993

I

n the past few years, numerous studies have systematically described the alterations in glucose metabolism that occur in NIDDM patients (1-3). These studies have found that in nonobese middle-aged NIDDM subjects with modest fasting hyperglycemia, basal HGO is increased despite fasting hyperinsulinemia, glucose-induced insulin release is impaired, and a marked resistance to insulin-mediated glucose disposal occurs. NIDDM is common in the elderly. By the age of 75, >20% of the U.S. population is afflicted with this disease (4). Despite the high prevalence of diabetes in this age-group, most studies of the alterations in glucose metabolism in NIDDM have excluded patients >65 yr of age (1-3). Normal aging is characterized by impaired glucose-induced insulin release (5) and resistance to insulinmediated glucose disposal (6,7), although these changes in CHO metabolism are not entirely attributable to aging itself, but are related in part to alterations in diet and physical activity (8). Therefore, older patients with NIDDM might be expected to have greater impairments in insulin release than middleaged NIDDM subjects because of the interaction of aging and diabetes. On the other hand, because normal aging already is characterized by insulin resistance, elderly NIDDM patients may have minimal insulin resistance when compared with age-matched control subjects. We conducted this study to examine alterations in glucose metabolism in the nonobese elderly with NIDDM. Untreated patients >65 yr of age with NIDDM were compared with normal control subjects carefully matched for age, sex, and obesity.

RESEARCH DESIGN AND METHODS — These studies were conducted in healthy, nonobese elderly subjects and elderly NIDDM patients. Normal healthy control subjects were recruited by advertising in senior citizens' publications. Subjects were ex-

1241

NIDDM in the elderly

eluded if they were 28 kg/m2, had evidence of underlying disease on history and physical examination, were taking medication, had abnormal laboratory tests, or had an abnormal glucose tolerance test based on the NDDG criteria (9). Of the initial responses to our advertisement, 70% were excluded, based on the above criteria. We recruited NIDDM patients by screening all patients seen in the Diabetes Centre during the year before the study who were >65 yr of age. Patients were excluded if they had a BMI >28 kg/m2 or had been treated with oral agents or insulin within the previous 3 mo. They also were excluded if they had evidence of complications from their diabetes or any other underlying disease, other than hypertension. Of the NIDDM patients, 2 were being treated for hypertension with calcium channel blockers; 4 patients had been treated in the past with glyburide, but in all cases the medication had been stopped because of frequent episodes of hypoglycemia, even on small doses of medication. In addition, 7 patients had a history of maturity-onset diabetes in a first-degree relative. All NIDDM patients had been diagnosed for 7.8 mM on two separate occasions and a diabetic OGTT based on the NDDG criteria. The patients enrolled in this study represented — 1% of our clinic population of elderly patients with NIDDM. The University of British Columbia Committee on Human Investigation approved this study. All participants gave written informed consent. Each study subject underwent an OGTT, a hyperglycemic glucose clamp, and a euglycemic glucose clamp. All consumed a diet containing >200 g of CHO for 3 days before the tests. Testing began at 0730, after a 12-h overnight fast in

1242

our Clinical Research Centre. Each test was separated by >2 wk. The dose of glucose in the OGTT was 40 g/m2. All participants remained recumbent during the test. Blood samples were taken at - 1 0 , 0, 30, 60, 90, 120, 150, and 180 min for measurement of glucose and insulin. Hyperglycemic and euglycemic glucose clamps were performed according to the method of Andres (10). In all studies, intravenous lines were inserted in an antecubital vein for infusion of glucose and in a contralateral hand vein for sampling of arterialized venous blood (11). In the hyperglycemic clamp, 4 blood samples were taken at 10-min intervals, from —30 to 0 min, to measure basal glucose and insulin values. At time 0, plasma glucose was acutely raised to 5.4 mM above basal and kept at that level for 120 min. During the study, glucose and insulin values were measured every 2 min from 0 to 10 min. Glucose was measured every 5 min and insulin every 10 min from 10 to 120 min. The CV of plasma glucose did not exceed 5% in any test. In the euglycemic clamp, R, and Rd were determined with primed constant infusions of 33H glucose (New England Nuclear, Boston, MA). All participants received a priming dose at —120 min, followed by a constant infusion until 240 min. In control subjects, the priming dose was 224 ± 2 nci/kg, ~100 times higher than that of the constant infusion rate of 2.33 ± 0.06 nci • kg" 1 • min" 1 . In normal subjects, a fixed priming dose 80—100 times greater than the constant infusion rate results in a steady state of tritiated glucose activity within 90 min (12). On the other hand, if a similar priming dose is given to patients with NIDDM, isotopic steady state will not be achieved for up to 360 min (12). Because of this, we increased the priming dose in patients with diabetes according to the formula: fasting glucose (mM) X calculated priming dose 5~mM

This method has been shown to allow isotopic steady state to be achieved within 60 min in diabetic patients (12). The average priming dose in our NIDDM patients was 403 ± 1 9 nci/kg, and the constant infusion was 2.22 ± 0.04 nci • kg" 1 • min" 1 . From —30 to 0 min, 4 blood samples were taken at 10-min intervals to measure basal glucose and insulin values. We took blood samples at the same time intervals to measure basal glucose specific activity to ensure that isotopic steady state was achieved before starting the clamp. At time 0, euglycemic clamp studies were started, which continued to 240 min. Regular human insulin (Humulin R, Lilly, Indianapolis, IN) was infused at a rate of 40 mU/m2 • min1. From 0 to 240 min, blood samples were taken every 5 min to measure glucose and every 15 min to measure insulin and glucose specific activity. In NIDDM patients, plasma glucose was allowed to fall to 5.5 mM before starting glucose infusion. Plasma glucose was maintained at this level for the rest of the study. The CV of plasma glucose did not exceed 5% in any study. Hother-Nielsen and Beck-Nieksen (13) showed that the use of a continuous infusion of tritiated glucose results in negative Rg values and an underrepresentation of GDRs, especially when the rates are high (13). Other studies have demonstrated this problem can be overcome if the cold glucose infusion is spiked with 33H glucose, the so-called HOT-GINF technique (13). We spiked our cold glucose infusion with 33H glucose according to the method of Finegood et al. (13). BP values were the mean of three values taken at 10-min intervals after the subjects had been recumbent for 30 min. WHR was calculated by dividing the largest abdominal girth by the hip circumference at the great trochanter. Bioelectric impedance was measured using a machine from RJL systems (Detroit, MI). The percentage of body fat was calculated from impedance measurements as previously described (14). Leisure-time

DIABETES CARE, VOLUME 16,

NUMBER 9, SEPTEMBER

1993

Meneilly, Dawson, and Tessier

Table 1—Clinical characteristics of study subjects ..-$

Normal subjects 9 73 ± 7/2 74.2 ± 25.7 ± 0.96 ± 33.4 ± 5453 ±

n Age (yr) Sex(M/F) Weight (kg) BM1 (Kg/M2) WHR Body fat (%) Physical activity (Kcal/wk) BP (mmHg) sBP dBP FBG (mM) HbA lc (%)f

1 2.0 0.4 0.01 3.2 1011

N1DDM patients 9 73 ± 2 7/2 77.2 ± 2.1 25.9 ± 0.5 1.00 ± 0.01* 29.1 ± 1.9 3301 ± 644

133 ± 4

144 ± 6

81 ±2 5.1 ±0.1

77 ± 3 9.3 ± 0.3t 8.1 ± 0.6



Data are means ± SE. *P < 0.05, normal subjects vs. diabetic patients. tP < 0.001, normal subjects vs. diabetic patients =FNormal range 4.4-6.4%.

physical activity was assessed with a questionnaire previously validated in large-scale epidemiological studies (15,16). Plasma glucose was measured immediately with the glucose oxidase method and a YSI glucose analyzer (Yellow Springs Instruments, Yellow Springs, OH). The remaining blood was placed in prechilled test tubes containing aprotonin (400 KlU/ml) and EDTA (1.5 mg/ml) and centrifuged at 4°C. The specific activity of glucose was determined from plasma samples deproteinized by barium hydroxide and zinc sulfate precipitation, as described previously (17). All RIA measurements were performed in duplicate as described previously (17). Our insulin assay cross-reacts —19% with proinsulin and its split products. All samples from an individual were analyzed at the same time, and included equal numbers of control subjects and patients with NIDDM in each assay. Rj and Rj were calculated with Steele's equations for non-steady-state conditions (18), modified for use with the HOT-GINF technique, as described by Finegood et al. (13). The volume of

DIABETES CARE, VOLUME 16,

,*''

Figure 1—Glucose (A) and insulin (B) values during the OGTT in normal control subjects (O) and NIDDM patients (9). Data are means ± SE.

distribution of glucose was assumed to be 210 ml/kg (19). All results are presented as means ± SE. The total AUC for insulin and glucose during the OGTT was calculated using the trapezoidal rule. Differences between groups were compared, using Student's t test for impaired samples. The glucose disposal data during the euglycemic clamp also were analyzed using repeated-measures ANOVA (P < 0.05 was considered significant). RESULTS— Table 1 presents characteristics of the study participants. Control subjects and NIDDM patients were similar in age, sex, weight, BMI, and BP. Although both groups had a similar percentage of body fat, the patients with NIDDM had a higher WHR. The control subjects appeared to be more active, but the difference did not reach statistical significance. The NIDDM patients, as expected, had an elevated fasting glucose and HbA lc . Figure 1 shows glucose and insulin values during the OGTT. In response to the OGTT, the AUC for glucose was greater in the NIDDM patients (16.0 ± 0.6 vs. 6.7 ± 0.3 mM in control subjects

NUMBER 9, SEPTEMBER

1993

vs. NIDDM patients, respectively, P < 0.001), and the AUC for insulin was less (238 ± 32 vs. 426 ± 53 pM in control subjects vs. NIDDM patients, respectively, P < 0.01). Figure 2 and Table 2 show glucose and insulin values during the hyperglycemic clamp. Fasting and steady-state (20-120 min) glucose values were higher during the hyperglycemic clamp in the patients with NIDDM. Fasting insulin was not significantly different between groups. The NIDDM patients had absent first-phase (0-10 min) insulin responses. Second-phase (20-120 min) insulin responses were substantially lower despite higher steady-state glucose values. Basal Rj values in control and patient groups are shown in Fig. 3. The distribution of values was similar in both groups. Glucose, insulin, glucose specific activity, and R^ and Rj during the euglycemic clamp are shown in Table 3 and Fig. 4. Although fasting glucose values were higher in the patients with NIDDM, glucose values were similar in both groups within 60 min and remained sim-

1243

A

20

i(mm

NIDDM in the elderly

10

•o u 3

5

5

Table 2—Glucose and insulin values during the hypergfycemic clamp Normal subjects

r

Fasting glucose (mM) 20-120 min glucose (mM) Fasting insulin (pM) 0-10 min insulin (pM) 20-120 min insulin (pM)

0

B

500

ulln( pmol/

400

M C

300

9 ±0.1 ±0.1 ±6 ± 31 ± 46

8.9 ± 0.3* 14.6 ± 0.4* 108 ± 6 112 ±6* 159 ± 11*

Data are means ± SE. *P < 0.01 normal subjects vs. diabetic patients.

200 100

5.4 11.0 96 250 337

NIDDM patients

**••* 40

60 80 Minute*

treated hypertension did not significantly CONCLUSIONS— In the past several alter the steady-state Rj in patients with years, alterations in glucose metabolism Figure 2—Glucose (A) and insulin (B) values NIDDM (steady-state Rj in nonhyperten- in NIDDM patients with varying degrees during the hypergfycemic clamp in normal con- sive patients: 34.1 ± 1 . 5 ixmol'kg" 1 ' of obesity have been extensively investitrol subjects (O) and patients with diabetes (•). min" 1 , NS). We found no differences gated (1-3). Numerous metabolic alterData are means ± SE. between groups in steady-state glucose ations have been described, including fasting hyperinsulinemia, impaired firstinfusion rates (Table 3). and second-phase insulin release, and Although steady-state GDRs were increased basal Rj. One of the most strikilar for the rest of the clamp. Fasting similar in the two groups, the pattern of ing metabolic abnormalities in NIDDM insulin was not significantly higher in the change over time in Rj appeared to be patients is a resistance to insulin-mediNIDDM patients. Insulin values during different. Accordingly, we analyzed the ated Rd (20-31), which is attributable to the clamp were equivalent. Specific ac- R^j data by repeated-measures ANOVA. defects in both oxidative and nonoxidativity values were relatively constant in We found a significant difference betive glucose disposal (1,2). Even in tween groups (F[l] = 6.74, P = 0.02), a both groups throughout the study. nonobese, middle-aged NIDDM patients Basal Rj, and Rj were similar in significant change over time in both with modest fasting hyperglycemia (FBG both groups. Steady-state (180-240 groups (F[16] = 85.65, P < 0.001), and